This evening, the 538 polling average moved about 2 points or so in the direction of the Mango Menace; at the time I write this it stands at a net -6.4% (dis)approval rating, with 51% disapproving and 44.6% approving. Yesterday the gap was -8. Two new polls seemed to affect the numbers—an Emerson College poll that had Trump at +4 (which 538 adjusted to even; this pollster apparently has a Republican lean), and an NBC News poll that was -3%. Of course, also dropping today were polls from Morning Consult and Yougov that had adjusted net approvals of -11% and -14%; but these pollsters (which poll the field more frequently, and are rated less highly by poblano and company) are given a much smaller weight in the polling average.
There hasn’t been any major news over the weekend (or past week) to explain a major jump, and the benefit of 538 is that it filters out a lot of noise, so when it DOES noticeably move, in either direction, I do take some note.
A couple possibilities:
- This is just polling noise/outlier polls. No amount of filtering will eliminate noise, and outliers are something to be expected statistically, so trying to exclude them is mathematically dubious. Last month, there was one poll that was -1%, which freaked everyone out, but it wasn’t reproduced. Since several of these polls are major media organizations which poll the field less frequently, they “move the needle” more.
- Response bias. While no major news has occurred that would move the undecided middle in any significant direction, the impeachment acquittal and subsequent “Trump unleashed” posture may have motivated his supporters, and caused them to be more willing to speak to pollsters, which can induce a bias in the polls. Reputable pollsters will publish results unfiltered, and also publish crosstrabs; 538 makes no attempt to “unskew” polls based on samples that differ from a statistical summary of the voter. (I looked at the crosstabs of the NBC poll, and did note that it appeared to be somewhat whiter and older than the population as a whole).
- Affects of impeachment. 538 did look into the question of how impeachment and acquittal affected things—while it didn’t seem to move opinion significantly, it may have soured people on the whole process (who view it as political). The nation remains split close to 50-50 on the topic, and I haven’t seen any evidence that Dems will be punished for it; but Trump and his media machine have been busy suggesting that a) it was just partisan politics, and b) what Trump did was OK.
- Democratic nominating contest. In case you haven’t noticed, the Dem nomination has moved from jawing to actual voting, and as usual, there’s been a lot of horserace coverage of that, in particular focusing on who might defect from the Democratic column if their favorite candidate doesn’t win. The GOP, on the other hand, is holding a coronation—having an incumbent, their candidate is not taking significant fire from inside the tent. Contested primaries in the opposition party can depress that party’s approval.
- Iowa caucus debacle. The Dems did step on their d*cks a few weeks ago (or the State of Iowa Democratic Party did), and the GOP got some good media coverage over that, but that seems to have faded from the news cycle of late.
- NH primary and the rise of Sanders. I don’t want to get into arguments of Who Can Best Beat Trump, nor suggest that any candidate should be cast aside because of general election electability concerns, but the past few weeks have also seen the collapse of the support of Joe Biden, and the rise of Senator Sanders to front-runner status. And… it seems that more than anyone else (except possibly Senator Warren, who seems to be being politely ignored, including by the same WSJ/NBC poll that pretended she doesn’t exist), a lot of powerful people in the news media Really Don’t Feel the Bern. (Said poll also contained much fear-mongering on the subject of socialism). Sanders still beats Trump in head-to-head polling, and national horserace polls are probably not terribly predictive (there are about 8-10 swing states that the election will come down to; and Sanders does better in some than others.
- Other news items: Coronavirus? Hasn’t been aggressively politicized, but I could see how a potential foreign-caused pandemic could aid the nationalist party and candidate—Ebola, far less of a threat to the US, was somewhat effectively politicized a while back. The economy? Trump gets good marks here; normally current economic conditions like this favor an incumbent, and there really hasn’t been any new news on this front.
- Polling conducted over a holiday weekend (President’s Day, no less)? Don’t discount this factor.
Thoughts?